DocumentCode
3546584
Title
Low complexity RDO mode decision based on a fast coding-bits estimation model for H.264/AVC
Author
Wang, Qiang ; Zhao, Debin ; Gao, Wen ; Ma, Siwei
Author_Institution
Dept. of Comput. Sci. & Technol., Harbin Inst. of Technol., China
fYear
2005
fDate
23-26 May 2005
Firstpage
3467
Abstract
Rate distortion optimization (RDO) is a very efficient tool for deciding a macroblock´s coding mode, which has been adopted by the H.264/AVC codec and brings higher coding efficiency. However, due to the precise calculation of the current macroblock´s distortion and coding-bits for each coding mode, RDO-based mode decision involves high computational complexity. In this paper, a simple model for quickly estimating the coding-bits of the quantized transform coefficients, which account for the major part of the computation when coding a macroblock, is presented. The model only uses nonzero coefficients´ magnitude and coordinates information by add operations to fulfill estimation so that computational complexity can be significantly reduced. Experimental results show that with this model the computational complexity of RDO-based mode decision in the H.264/AVC codec can be reduced by 40-60% while coding efficiency has little drop.
Keywords
computational complexity; video coding; H.264/AVC; coding efficiency; computational complexity reduction; fast coding-bits estimation model; low complexity RDO-based mode decision scheme; macroblock coding mode decision; macroblock-based multimode coding scheme; nonzero coefficient coordinates information; nonzero coefficient magnitude information; quantized transform coefficients; rate distortion optimization; Automatic voltage control; Bit rate; Codecs; Computational complexity; Computational modeling; Computer science; Cost function; Motion estimation; Rate-distortion; Video coding;
fLanguage
English
Publisher
ieee
Conference_Titel
Circuits and Systems, 2005. ISCAS 2005. IEEE International Symposium on
Print_ISBN
0-7803-8834-8
Type
conf
DOI
10.1109/ISCAS.2005.1465375
Filename
1465375
Link To Document